The Epigenetics Revolution

Chapter 14. Long Live the Queen

The effects of nutrition on the health and lifespan of mammals are pretty dramatic. As we saw in the previous chapter, prolonged calorie restriction can extend lifespan by as much as one-third in mice[261]. We also saw in Chapter 6 that our own health and longevity can be affected by the ways our parents and grandparents ate. These are quite startling findings but nature has provided us with a much more dramatic example of the impact of nutrition on lifespan. Imagine, if you can, a dietary regime that means a select few in a species have a lifespan that is twenty times longer than that of most of their companions. Twenty times longer. If that happened in humans, the UK might still be in the reign of Queen Elizabeth I, and would expect to be so for about another 400 years.

Obviously this doesn’t happen in humans, but it does happen in one common organism. It’s a creature that we all meet every spring and summer. We use the results of its labour to make candles and furniture polish, and we have eaten its hard-earned bounty since the very beginning of human history. It’s the honeybee.

The honeybee, Apis mellifera, is a truly extraordinary creature. It is a prime example of a social insect. It lives in colonies that can contain tens of thousands of individuals. The vast majority of these are workers. These are sterile females, which have a range of specialised roles including gathering pollen, building living quarters and looking after the young. There are a small number of males, who do very little except mate, if they are lucky. And there is a queen.

In the formation of a new colony, a virgin queen leaves a hive, accompanied by a swarm of workers. She’ll mate with some males and then settle down to form a new colony. The queen will lay thousands of eggs, most of which will hatch and develop into more workers. A few eggs will hatch and develop into new queens, who can start the whole cycle all over again.

Because the queen who founded the colony mated several times, not all the bees in the colony will be genetically identical to each other, because some of them will have different fathers. But there will be groups of thousands and thousands of genetically identical bees in any colony. This genetic identity doesn’t refer only to the worker bees. The new queens are genetically identical to thousands of worker bees in the colony. We could call them sisters, but this doesn’t really describe them well enough. They are all clones.

However, a new queen and her clonal worker sisters are clearly incredibly different from each other, both in physical form and in activities. The queen can be twice the size of a worker bee. After the so-called nuptial flight, when she first leaves a colony and mates, the queen almost never leaves the hive again. She stays in the darkness of the interior, laying up to 2,000 eggs a day in the summer months. She has no sting barbs, no wax glands and no pollen baskets (not much point having a shopping bag if you never leave the house). Worker bees have a lifespan that can usually be measured in weeks, whereas queens live for years[262].

Conversely, workers can do many things that the queens can’t. Chief amongst these is collecting food, and then telling the rest of the colony its location. This information is communicated using the famous ‘waggle dance’. The queen lives in darkened luxury, but she never gets to boogie.

So, a honeybee colony contains thousands of individuals who are genetically identical, but a few of them are really different physically and behaviourally. This difference in outcome is all down to how the bee larvae are fed. The pattern of early feeding completely determines whether a larva will develop into a worker or into a queen.

For honeybees the DNA script is constant but the outcome is variable. The outcome is controlled by an early event (feeding pattern) which sets a phenotype that is maintained throughout the rest of life. This is a scenario that just shrieks epigenetics at us, and in the last few years scientists have started to unravel the molecular events that underpin this process.

The critical roll of the dice for honeybees happens after the third day of life, as a fairly immobile grub or larva. Until day three, all honeybee larvae are given the same food. This is a substance called royal jelly, which is produced by a specialised group of workers. These young workers are known as nurse bees and they secrete royal jelly from glands in their heads. Royal jelly is a highly nutritious food source. It is a concentrated mix of a lot of different components, including key amino acids, unusual fats, specific proteins, vitamins and other nutrients that haven’t been well-characterised yet.

Once the larvae are three days old, the nurse bees stop feeding royal jelly to most of them. Instead, most larvae are switched onto a diet of pollen and nectar. These are the larvae which will grow up to be worker bees.

But for reasons that nobody really understands, the nurse bees continue feeding royal jelly to a select few larvae. We don’t know how these larvae are chosen or why. Genetically they are identical to the ones that are switched onto the less sophisticated diet. But this small group of larvae that continue to be nourished with royal jelly grow up to be queens, and they’re fed this same substance throughout their lives. The royal jelly is essential for the production of mature ovaries in the queens. Worker females never develop proper ovaries, which is one of the reasons they are infertile. Royal jelly also prevents the queen from developing the organs that she won’t ever need, like those pollen baskets.

We understand some of the mechanisms behind this process. Bee larvae contain an organ that has some of the same functions as our liver. When a larva receives royal jelly continuously, this organ processes the complex food source and activates the insulin pathway. This is very similar to the hormonal pathway in mammals that controls the levels of sugar in the bloodstream. In honeybees activation of this pathway increases production of another hormone, called Juvenile Hormone. Juvenile Hormone in turn activates other pathways. Some of these stimulate growth and development of tissues like the maturing ovaries. Others shut down production of the organs that the queen doesn’t need[263].

Because honeybee maturation has so many hallmarks of an epigenetic phenomenon, researchers speculated that there would also be an involvement of the epigenetic machinery. The first indications that this is indeed the case came in 2006. This was the year when researchers sequenced the genome of this species, to identify the fundamental genetic blueprint[264]. Their research showed that the honeybee genome contained genes that looked very similar to the DNA methyltransferase genes of higher organisms such as vertebrates. The honeybee genome was also shown to contain a lot of CpG motifs. This is the two-nucleotide sequence that is usually the target for DNA methyltransferases.

In the same year, a group led by Gene Robinson in Illinois showed that the predicted DNA methyltransferase proteins encoded in the honeybee genome were active. The proteins were able to add methyl groups to the cytosine residue in a CpG motif in DNA[265]. Honeybees also expressed proteins that were able to bind to methylated DNA. Together, these data showed that honeybee cells could both ‘write’ and ‘read’ an epigenetic code.

Until these data were published, nobody had really wanted to take a guess as to whether or not honeybees would possess a DNA methylation system. This was because the most widely used experimental system in insects, the fruit fly Drosophila melanogaster, whom we met earlier in this book, doesn’t methylate its DNA.

It’s interesting to discover that honeybees have an intact DNA methylation system. But this doesn’t prove that DNA methylation is involved in the responses to royal jelly, or the persistent effects of this foodstuff on the physical form and behaviour of mature bees. This issue was addressed by some elegant work from the laboratory of Dr Ryszard Maleszka at the Australian National University in Canberra.

Dr Maleszka and his colleagues knocked down the expression of one of the DNA methyltransferases in honeybee larvae, by switching off the Dnmt3 gene. Dnmt3 is responsible for adding methyl groups to regions of DNA that haven’t been methylated before. The results of this experiment are shown in Figure 14.1.

^{Figure 14.1 When royal jelly is fed to honeybee larvae for extended periods, the larvae develop into queens. The same effect is seen in the absence of prolonged feeding with royal jelly if the expression of the Dnmt3 gene is decreased experimentally in the larvae. Dnmt3 protein adds methyl groups to DNA.}

When the scientists decreased the expression of Dnmt3 in the honeybee larvae, the results were the same as if they had fed them royal jelly. Most of the larvae matured as queens, rather than as workers. Because knocking down Dnmt3 had the same effects as feeding royal jelly, this suggested that one of the major effects of royal jelly is connected with alterations of the DNA methylation patterns on important genes[266].

To back up this hypothesis, the researchers also examined the actual DNA methylation and gene expression patterns in the different experimental groups of bees. They showed that the brains of queens and worker bees have a different DNA methylation pattern. The DNA methylation patterns in the bees where Dnmt3 had been knocked down were like those of the normal royal jelly-induced queens. This is what we would expect given the similar phenotypes in the two groups. The gene expression patterns in the normal queens and the Dnmt3-knockdown queens were also very similar. The authors concluded that the nutritional effects of continual feeding on royal jelly occurred via DNA methylation.

There are still a lot of gaps in our understanding of how nutrition in the honeybee larva results in altered patterns of DNA methylation. One hypothesis, based on the experiments above, is that royal jelly inhibits the DNA methyltransferase enzyme. But so far nobody has been able to demonstrate this experimentally. It’s therefore possible that the effect of royal jelly on DNA methylation is indirect.

What we do know is that royal jelly influences hormonal signalling in honeybees, and that this changes gene expression patterns. Changes in the levels of expression of a gene often influence the epigenetic modifications at that gene. The more highly a gene is switched on, the more its histones become modified in ways which promote gene expression. Something similar may happen in honeybees.

We also know that the DNA methylation systems and histone modification systems often work together. This has created interest in the role of histone-modifying enzymes in the control of honeybee development and activity. When the honeybee genome was sequenced, four histone deacetylase enzymes were identified. This was intriguing because it has been known for some time that royal jelly contains a compound called phenyl butyrate[267]. This very small molecule can inhibit histone deacetylases but it does so rather weakly. In 2011, a group led by Dr Mark Bedford from the MD Anderson Cancer Center in Houston published an intriguing study on another component of royal jelly. One of the other senior authors on this paper was Professor Jean-Pierre Issa, who has been so influential in promoting use of epigenetic drugs for the treatment of cancer.

The researchers analysed a compound found in royal jelly called (E)-10-hydroxy-2-decenoic acid, or 10HDA for short. The structure of this compound is shown in Figure 14.2, along with SAHA, the histone deacetylase inhibitor we saw in Chapter 11 that is licensed for the treatment of cancer.

^{Figure 14.2 The chemical structure of the histone deacetylase inhibitor SAHA and 10HDA, a compound found in royal jelly. C: carbon; H: hydrogen; N: nitrogen; O: oxygen. For simplicity, some carbon atoms have not been explicitly shown, but are present where there is a junction of two lines.}

The two structures aren’t identical by any means, but they do share some similarities. Each has a long chain of carbon atoms (the bit that looks like a crocodile’s back in profile), and the right hand side of each compound also looks fairly similar. Mark Bedford and his colleagues hypothesised that 10HDA might be an inhibitor of histone deacetylases. They performed a number of test tube and cell experiments and showed that this was indeed the case. This means that we now know that one of the major compounds found in royal jelly inhibits a key class of epigenetic enzymes[268].

Epigenetics influences more than whether bees develop into workers or queens. Ryszard Maleszka has also shown that DNA methylation is involved in how honeybees process memories. When honeybees find a good source of pollen or nectar, they fly back to the hive and tell the other members of the colony where to head to find this great food supply. This tells us something really important about honeybees; they can remember information. They must be able to, or they wouldn’t be able to tell the other bees where to go for food. Of course, it’s equally important that the bees can forget information and replace it with new data. There’s no point sending your co-workers to the great patch of thistles that were in flower last week, but that have now been eaten by a grazing donkey. The bee needs to forget last week’s thistles and remember the location of this week’s lavender.

It’s actually possible to train honeybees to respond to certain stimuli associated with food. Dr Maleszka and his colleagues showed that when the bees undergo this training, the levels of Dnmt3 protein go up in the parts of the honeybee brains which are important in learning. If the bees are treated with a drug that inhibits the Dnmt3 protein, the compound alters the way the bees retain memories, and also the speed with which memories are lost[269].

Although we know that DNA methylation is important in honeybee memory, we don’t know exactly how this works. This is because it’s not clear yet which genes become methylated when honeybees learn and acquire new memories.

So far, we could be forgiven for thinking that honeybees and higher organisms, including us and our mammalian relatives, all use DNA methylation in the same way. It’s certainly true that changes in DNA methylation are associated with alterations in developmental processes in both humans and honeybees. It’s also true that mammals and honeybees both use DNA methylation in the brain during memory processing.

But oddly enough, honeybees and mammals use DNA methylation in very different ways. A carpenter has a saw in his toolbox and uses it to build a book case. An orthopaedic surgeon has a saw on his operating trolley and uses it to amputate a leg. Sometimes, the same bit of kit can be used in very different ways. Mammals and honeybees both use DNA methylation as a tool, but during the course of evolution they’ve employed it very differently.

When mammals methylate DNA, they usually methylate the promoter regions of genes, and not the parts that code for amino acids. Mammals also methylate repetitive DNA elements and transposons, as we saw in Emma Whitelaw’s work in Chapter 5. DNA methylation in mammals tends to be associated with switching off gene expression and shutting down dangerous elements like transposons that might otherwise cause problems in our genomes.

Honeybees use DNA methylation in a completely different way. They don’t methylate repetitive regions or transposons, so they presumably have other ways of controlling these potentially troublesome elements. They methylate CpG motifs in the stretches of genes that encode amino acids, rather than in the promoter regions of genes. Honeybees don’t use DNA methylation to switch off genes. In honeybees, DNA methylation is found on genes that are expressed in all tissues, and also on genes that tend to be expressed by many different insect species. DNA methylation acts as a fine-tuning mechanism in honeybee tissues. It modulates the activity of genes, turning the volume slightly up or down, rather than acting as an on-off switch[270]. Patterns of DNA methylation are also strongly correlated with control of mRNA splicing in honeybee tissues. However, we don’t yet know how this epigenetic modification actually influences the way in which a message is processed[271].

We’re really only just beginning to unravel the subtleties of epigenetic regulation in honeybees. For example, there are 10,000,000 CpG sites in the honeybee genome, but less than 1 per cent of these are methylated in any given tissue. Unfortunately, this low degree of methylation makes analysing the effects of this epigenetic modification very challenging. The effects of Dnmt3 knockdown show that DNA methylation is very important in honeybee development. But, given that DNA methylation is a fine-tuning mechanism in this species, it’s likely that Dnmt3 knockdown results in a number of individually minor changes in a relatively large number of genes, rather than dramatic changes in a few. These types of subtle alterations are the most difficult to analyse, and to investigate experimentally.

Honeybees aren’t the only insect species that has developed a complex society with differing forms and functions for genetically identical individuals. This model has evolved independently several times, including in different species of wasps, termites, bees and ants. We don’t yet know if the same epigenetic processes are used in all these cases. Shelley Berger from the University of Pennsylvania, whose work on ageing we encountered in Chapter 13, is involved in a large collaboration focusing on ant genetics and epigenetics. This work has already shown that at least two species of ants also can methylate the DNA in their genomes. The expression of different epigenetic enzymes varies between different social groups in the colonies[272]. These data tentatively suggest that epigenetic control of colony members may prove to be a mechanism that has evolved more than once in the social insects.

For now, however, most interest in the world outside epigenetics labs focuses on royal jelly, as this has a long history as a health supplement. It’s worth pointing out that there’s very little hard evidence to support this having any major effects in humans. The 10HDA, that Mark Bedford and his colleagues showed was a histone deacetylase inhibitor, can affect the growth of blood vessel cells[273]. Theoretically, this could be useful in cancer, as tumours rely on a good blood supply for continuing growth. However, we’re a very long way from showing that royal jelly can really fight off cancer, or aid human health in any other way. If there’s one thing we do already know, it’s that bees and humans are not the same epigenetically. Which is just as well, unless you’re a really big fan of the monarchy …